NVIDIA Broadcast transforms any space into a home broadcast studio, upgrading standard webcams and microphones into premium, smart devices through the power of AI. Improve your livestream video and audio quality with AI capabilities such as virtual background, webcam auto-framing, and microphone noise removal. With dedicated AI processors called Tensor Cores on GeForce RTX GPUs, AI networks can run high-quality effects in real time.
When not broadcasting, these features can also enhance your video conference calls and voice chats, making NVIDIA Broadcast a perfect AI companion for anyone.
The NVIDIA Broadcast app includes:
Microphone/Speaker
Camera
Yes! In version 2.0.0 we introduced the ability to run more than one effect at a time. You may combine as many effects in your microphone, speaker and/or camera, as your GPU can handle. Not all effects can be combined.
Please note that combining effects increases impact on your system.
Video Noise Removal is an AI effect that reduces visual noise from your camera feed, which typically occurs in low-light situations. This effect is useful for webcams, as their sensors oftentimes capture visual noise, giving off the appearance of a low-quality feed. Video Noise Removal cleans up the visual noise, providing higher picture quality akin to a premium webcam.
Room Echo Removal is an AI effect that reduces the room echo (or reverb) from your audio. This typically happens in spaces without sound isolation. This is especially useful for users working or studying from home in smaller rooms without proper soundproofing.
RTX Video Super Resolution upscales webcam and file video up to 4K resolution. In live mode, choose between Performance and Quality modes. Only resolutions higher than the source are offered. Requires a GeForce RTX 3060 or higher.
RTX Video Frame Generation increases webcam frame rate up to 60 fps in live mode, and increases file video frame rate by 2x or 4x. In live mode, choose between Performance and Balanced modes. Requires a GeForce RTX 40 series GPU or higher.
Studio Voice offers four mic profiles — Default, Bright, Full, and Warm — that adjust the tonal character of your voice. Default applies no additional coloration. Bright sounds clear and sharp. Full sounds rich and deep. Warm sounds soft, smooth, and bass-forward.
Files mode lets you apply supported NVIDIA Broadcast audio and video effects to recorded media files. Supported formats are standard .mp4, .m4a, .mp3, and .wav. Supports .h264 (8bit) and aac codecs and up to 2 audio channels. Add files, preview input, choose effects, and process them through a queue. Since this is an offline mode, some effects can take longer to deliver higher quality than real-time mode.
Currently, File Processor does not support the fragmented MP4 format. Use or convert to Standard MPEG-4 mp4 files before using them in the File Processor.
NVIDIA Broadcast includes a local MCP (Model Context Protocol) server that lets compatible AI agents on the same computer read and change supported settings. Find connection information under Settings → AI Agent Resources. The connection is local only — no external access is possible.
Agents can operate the following features in NVIDIA Broadcast through the MCP server:
The following isn’t supported: Settings page settings, test and preview capabilities for camera and audio effects
NVIDIA Broadcast works with most PC microphones, speakers or headsets, and webcams. These include most recent models from popular manufacturers such as Logitech, Elgato, AverMedia, Razer, Asus, HyperX, SteelSeries, Rode and Bose.
Make sure that if your device has a similar feature to NVIDIA Broadcast (e.g. Noise Removal) you turn it off, so you don’t try to apply the same effect twice.
NVIDIA Broadcast is compatible with a wide range of broadcasting, video conferencing and voice chat apps. These include: OBS Studio, Streamlabs, Xsplit Broadcaster, HuYa, DouYu, Bilibili, Discord, TeamSpeak, Skype, Zoom, Webex, Microsoft Teams, Slack, Google Meet, and more.
NVIDIA Broadcast supports keyboard shortcuts. Optionally, users can also use the Elgato Stream Deck to turn on and off the NVIDIA Broadcast effects. If you are interested in the Elgato Stream Deck plugin, visit the Elgato Marketplace on this link.
NVIDIA Broadcast works by creating a virtual Windows camera, microphone, or speaker on your PC, so it appears to video or voice chat applications like a physical device. Simply enable features from within the NVIDIA Broadcast app, then select the NVIDIA Broadcast device from your video or voice chat application.
Virtual Key Light, Studio Voice, and RTX Video Super Resolution (beta) require a GeForce RTX 3060 GPU or higher. RTX video Frame Generation (beta) requires a GeForce RTX 40 series GPU or newer.
During the installation process, NVIDIA Broadcast will download the latest AI models for your GPU. This ensures that you have the latest version, and reduces the download size by only requiring you to download the models needed for your GPU.
No. The NVIDIA Broadcast app uses Tensor Cores on NVIDIA RTX GPUs to accelerate AI calculations so you can game, livestream and run AI networks at the same time.
For those of you that want to try out the AI noise removal capabilities but aren’t ready to upgrade to an RTX GPU yet, we are also patching RTX Voice with support for NVIDIA GeForce GTX GPUs. Though your mileage may vary on older cards.
You can download NVIDIA Broadcast here.
NVIDIA Broadcast includes noise removal capabilities based on the RTX Voice community beta and more. It also includes AI-assisted virtual background and auto frame features.
The performance impact depends on the GPU, game, webcam resolution and frame rate, and the number of effects used during the broadcast. But in general, we’ve measured the performance impact of the effects to be within 5% and 10% on a GeForce RTX 5090
Studio Voice is designed to push your mic quality to the next level and is intended for non-gaming livestreaming (just chatting), video conferencing, or content creation. It uses a larger portion of your GPU, so we don't recommend using it while gaming or running a GPU-intensive application.
If you experience performance issues, we recommend turning off effects you are not using and reducing the camera resolution and/or frame rate.
On the NVIDIA Broadcast app, click on the “Feedback” button in the upper left corner of the app
You can get support here: broadcast.nvidia.com/support
You can help us train the AI here: broadcast.nvidia.com/feedback
Developers interested in integrating the technology of NVIDIA Broadcast into their application can get access to the NVIDIA AI for Media SDKs from our developer site: developer.nvidia.com/topics/ai/generative-ai/ai-for-media.
A GeForce RTX 4080, 5070, or higher.
The Studio Voice effect is suitable for non-gaming use cases (just chatting, livestreams, video conferencing, and content creation).